长江干流站中长期径流预报方法研究
Research on Medium and Long Term Runoff Forecast in Yangtze River Basin

DOI: 10.12677/JWRR.2014.34035, PP. 283-290

Keywords: 逐步回归，支持向量机，长江流域，中长期径流预报
Stepwise Regression, Support Vector Machine, Yangtze River Basin, Medium and Long-Term Runoff Forecast

Full-Text   Cite this paper   Add to My Lib

Abstract:

长江流域是我国经济最发达的地区之一，水资源量的多少直接影响着该地区经济社会的发展，因此准确的中长期预报对于水库群联合调度、水资源调配和合理利用具有重要的意义。本文以长江流域屏山、宜昌、大通和汉口4个站点为研究对象，依据74项大气环流指数和前期径流共75项预报因子，采用相关系数法初选及逐步回归法优选预报因子，建立了基于支持向量机的月、旬径流预报模型，定量分析了模型在长江流域的适用性，并与人工神经网络预测模型进行比较，结果表明该模型在长江流域中长期径流月、旬预报中，检验期平均合格率分别为49.29%和54.49%，达不到实际应用的需求，而旬尺度计算月径流的结果优于月尺度的模拟结果，有较好的预报精度，可为长江流域水文预报工作提供参考。相对而言，支持向量机模型的预测精度优于人工神经网络模型。
Yangtze River is one of China’s most economically developed regions, and the quantity of water directly affects the economic and social development in the region, so an accurate medium and long-term forecast is significant for multi-reservoir scheduling, water resources allocation and ra-tional utilization. In this paper, Pingshan, Yichang, Datong and Hankou four stations in the Yangtze River basin were selected as the research objects, based on 75 predictors, including 74 atmospheric circulation index and pre-runoff, using the correlation coefficient method to preliminary select the predictors and stepwise regression method to optimize the predictors, a runoff forecasting model of monthly scale and ten-day scale based on the Support vector machine (SVM) was established. And the applicability of the model in the Yangtze River basin was quantitatively analyzed. In addition, the prediction precision of the model was compared with that of artificial neural network prediction model. The results indicate that the model based on SVM can’t meet the actual application requirements, because the average qualified rates during the test period were only respectively 49.29% and 54.49% in the medium and long-term runoff forecasting of monthly scale and ten-day scale. But the monthly runoff results calculated by ten-day scale are superior to those by monthly scale. The former can provide a reference for the work of the Yangtze River hydrological forecasting. Relatively speaking, the prediction precision of the model based on Support vector machine (SVM) is better than that of artificial neural network model.

References

[1]	李红波, 夏潮军, 王淑英. 中长期径流预报研究进展及发展趋势[J]. 人民黄河, 2012, 34(8): 36-40. LI Hongpo, XIA Chaojun and WANG Shuying. Research progress and development trend of medium-long term runoff forecast. Yellow River, 2012, 34(8): 36-40. (in Chinese)
[2]	张兰影, 庞博, 徐宗学. 基于支持向量机的石羊河流域径流模拟适用性评价[J]. 干旱区资源与环境, 2013, 27(7): 113-118. ZHANG Lanying, PANG Bo and XU Zongxue. Assessment on the applicability of support vector machine-based models for runoff simulation in Shiyang river basin. Journal of Arid Land Resources and Environment, 2013, 27(7): 113-118. (in Chinese)
[3]	BOX, G. E. P., JENKINS, G. M. Time series analysis forecasting and control. San Fran-cisco: Holden-Day, 1976.
[4]	王伟, 雷晓辉. 石羊河流域中长期径流预报模型的应用[J]. 人民黄河, 2014, 36(1): 42-44. WANG Wei, LEI Xiaohui. Research on medium and long term hydrological statistics prediction in runoff forecasting in the Shiyang River Basin. Yellow River, 2014, 36(1): 42-44. (in Chinese)
[5]	李继伟, 纪昌明. 基于支持向量机的水电站中长期径流组合预报[J]. 水电能源科学, 2013, 31(11): 13-16. LI Jiwei, JI Changming. Medium and long-term runoff combination forecast based on support vector machine. Water Resources and Power, 2013, 31(11): 13-16. (in Chinese)
[6]	崔东文, 郭荣. 基于几种参数优化的支持向量机在径流预报中的比较分析[J]. 水资源研究, 2013, 34(2): 34-38. CUI Dongwen, GUO Rong. Supportvector machine (SVM) in the comparative analysis of runoff forecast based on several parameters optimization. Journal of Water Resources Research, 2013, 34(2): 34-38. (in Chinese)
[7]	郭俊, 周建中. 基于改进支持向量机回归的日径流预测模型[J]. 水力发电, 2013, 36(3): 12-15. GUO Jun, ZHOU Jianzhong. Daily runoff forecast based on improved support vector machine regression model. Water Power, 2013, 36(3): 12-15. (in Chinese)
[8]	王双银, 杨筱筱. 基于相关系数和Fisher最优分割法的汛期分期研究[J]. 水文, 2011, 32(4): 1-5. WANG Shuangyin, YANG Xiaoxiao. Flood season division based on correlation coefficient and optimum partition method of fisher. Journal of China Hydrology, 2011, 32(4): 1-5. (in Chinese)
[9]	李琪, 马建斌, 刘洪吉. 基于逐步回归的投影寻踪水文预报模型研究[J]. 水电能源科学, 2011, 29(2): 10-12. LI Qi, MA Jianbin and LIU Hongji. Research on projection pursuit forecast model based on stepwise regression. Water Resources and Power, 2011, 29(2): 10-12. (in Chinese)
[10]	宋星原, 张利平, 张艳军. 长江流域水资源预测预报方案编制研究工作[R]. 武汉: 长江水利委员会水文局, 2014. SONG Xingyuan, ZHANG Liping and ZHANG Yanjun. Water resources prediction scheme establishment research work in the Yangtze River Basin. Wuhan: Water Resources Commission Hydrology Bureau of Yangtze River, 2014. (in Chinese)

Full-Text